A nice dictionary, but how can I find what I want?

-- On the Longman Language Activator

TOMIOKA, Taeko

Associate Professor, Sanno College, Jiyugaoka

Abstract:

A few years have passed since Longman's Language Activator (henceforward Activator) was published as "the world's first production dictionary". In spite of the fact that this dictionary is filled with natural and authentic example sentences, inspiring definitions and a number of clever features to help students striving to convert meanings into words, it still doesn't seem to have become very popular among Japanese students. In this paper, I will show how difficult it often is to get to what you want to say, especially for Japanese students of English. Then I will suggest how it could be improved.

Introduction

When Activator was published, everyone was surprised by its unprecedented attempt to help students express natural English. This dictionary was created based on a thorough study of various corpora, including second language learners' corpus. Careful attention was paid to choosing only useful and common expressions for the production part of language activities. As is proudly stated in the introduction, this is an "encoding" dictionary, not a "decoding" dictionary. After browsing through this dictionary more carefully, we were even more impressed with its striking features. It is filled with useful expressions, authentic example sentences, and explanations to help students understand the differences among several expressions and their usage. Since this is fully described in Miyai et al. (1995) and Tomioka (1996), I'll not repeat it here.

The problem arises when students have this dictionary in front of them, have some idea they want to put into English and they don't easily succeed in finding the expressions they want. I'll discuss below several problems in the look-up system of this dictionary, show how it is actually difficult for Japanese students to find the expressions they want, based on an experiment I conducted, and then suggest some possible solutions to these problems.

Problems

Since this dictionary is completely written in English, users should start with the Key Word which they think represents or is related to the ideas they want to express. For example, if you are looking for the expression meaning "to walk very softly and stealthily on the tips of the toes, not to be heard," you have to conclude that it is a manner of walking, so it should be under the Key Word WALK. Sure enough, as is shown in the Fig.1 below, it leads to the idea "to walk quietly with light steps" and then to the word "tiptoe" along with several other expressions like "pad," "creep" or "sneak." as in Fig.2.

In this dictionary ideas are grouped into 1052 semantic concepts called Key Words so that users can find the expressions only from the meaning. In case users cannot think of or find an appropriate Key Word but they know a similar or related word to the idea, they can look for the words in a list of headwords, which is also alphabetically presented in the body of the dictionary. The list itself doesn't give any definition or explanation, but it leads to likely Key Words. In this case, if the user cannot think of the Key Word, WALK, but does happen to know the expression "sneak", then he can start there.

The question is whether this look-up system is efficient and friendly enough for the user. In other words, can the users easily think of an appropriate Key Word and that without taking too long?

One example shows there is a problem. If the user is looking for the expression meaning "squeeze the windpipe or neck of somebody, esp. so as to kill," or "to kill someone by pressing on their throat with both hands or with something such as a piece of string," what is the appropriate Key Word for that? "squeeze", "neck", "kill" "murder" or "throat"? "squeeze" is found in the list of headwords and it leads to several Key Words as in Fig.3, but none of them leads to the expression you want.

"kill" is a Key Word and as is shown in Fig.4, it does lead to several ways of killing a person, including capital punishment, euthanasia or suicide, but not the expression "strangle."

In this case, the appropriate Key Word was "breathe". Key Word BREATHE has several expressions like "choke" "suffocate" "strangle" and "smother" in its tenth meaning -- "to kill someone by stopping them from breathing." It is a little surprising and misleading not only for students but also for us English teachers that some manners of killing are grouped under KILL, while others under BREATHE. Users try all the possibilities above, and yet still find nothing.

Another problem is that similar expressions are under different Key Words and there is no way to go one from another unless we think of other appropriate Key Words. For example, we would think the expression "to finish/end feebly" is under the Key Word FINISH/COME TO AN END (see Fig.5), because it is a manner of something coming to an end.

But there we can only find the expressions meaning "finish/end" and the meaning we are looking for is, in fact, under STOP HAPPENING as is shown in Fig.6.

You need to think of that Key Word to find some interesting expressions such as fizzle out, peter out, or wear off. Also a similar expression die away is under DISAPPEAR. (see Fig.7)

These words scattered under separate Key Words have very close meanings and it would be very useful for the users to be able to consider other possibilities, but there is no device in this dictionary to retrieve expressions similar to each other.

The last problem I'd like to present here is how many times the students have to turn pages to get to what they want. Unlike regular Japanese-English dictionaries, where you just go to Japanese headwords and the translations of those words are given, you have to first go to the Key Word which you think is appropriate. Out of the several meanings given there, you then choose the one close to your idea, and then go to that section. There is, therefore, always one step more to take even if you can decide on the best Key Word. If the key Word you first thought of doesn't work, then you have to double or triple these steps. The more the students have to turn pages, the more frustrating and less likable it will become to use this dictionary.

It is also really a shame not to be able to access such wonderful information Activator contains.

The Experiment

I conducted an experiment in order to see how easy or difficult it is for Japanese students to access the expressions they want. A questionnaire which looks like Fig.8 was given to 35 junior college students (30 female and 5 male) at Sanno. They are between age 18 and 20. Their level of English is between low-intermediate and intermediate. None of them had used this dictionary before, nor had they known about this dictionary.

After a full explanation of the Key Word system using real examples, the students were asked to write down possible words or expressions they might use as a Key Word or they might look up in the list of headwords. The meanings were given in Japanese words. They had a 1052 Key Word list copied from the back part of Activator, but they didn't have an actual dictionary at hand. They were allowed to give as many as four candidates, but they were asked to write them in the order that they would look up. In case they could not come up with any appropriate English word, they were allowed to answer in Japanese. If they happened to know the exact expression for the meaning, they were also allowed to write it down along with other possible Key Words. Then their answers were scored depending on whether they would actually lead to the desired expressions.

Discussions

As is shown in Fig.9, very few students successfully found the expressions. The most difficult ones were "prosecution" and "temporary," and only 6 to 8% succeeded, although for "talkative" and "burn," about 60 to 70% did.

Table 1 below shows the overall performance of each student, in other words, out of ten meanings they were given, how many they got as their first choice - only a little over two. Even after adding their second to fourth choices, their success rate is only 37%. If they can find what they want only four times out of ten and that with lots of turning pages, you cannot blame the students for not wanting to use that dictionary.

First choice
Other choices
Total

average/10
2.286
1.429
3.771

Table 1 Average scores of finding the words

Table 2 shows all the words the students thought appropriate for the Key Words or look-up words. The numbers on the left of the words show how many students chose. The shadowed words are the ones which successfully led to the expressions and the numbers on the right in circle show how many times they had to turn pages to get to the expressions.

Table 2 suggests several problems Activator's look-up system has:

2) There are some very language specific associations the students made.

3) You sometimes need to know the exact word to get to the expressions, which spoils the purpose of the dictionary. [Examples] "temporary" in #1, "cure/treat" in #2, etc.

4) Even if you know words very close to the meaning, this look-up system sometimes fails to lead the user to the expressions. [Examples] "fever" in #4, "crazy" in #5 "weight" in #8, etc.

Since the third and fourth problems are very obvious, let me talk about the first two problems. "talkative" in #9 and "prolific" in #10 are each likely to be grouped under "talk," "much/a lot" and "produce/make," "much/a lot" respectively. But actually, "talkative" can be found from the words talk, speak, say, but not from much/a lot, while "prolific" can be only found from much/a lot, not from produce/make. In the same way, "submerge" is searchable from the Key Word DOWN, but "lose weight" is not, even if you can think a person's weight "goes down". The Key Word DOWN has various meanings concerning several things going down; such as the sun, an airplane, a person, a car, etc. It also includes something with less physical movement such as demote, downgrade, but it does not include "The airplane crashes," "The stock market slumps," or "The company goes bankrupt." This shows how the semantic categorization is often very unpredictable and it takes a lot of trial and error to get to what you want.

The second problem is that some associations the students made seem very language specific. For example, because of Kanji compounds muchuu (dream+in) -- "be crazy about" and sensui (dive+water) -- "submerge," or ichijitekina (one+time) -- "temporary," 60 to 75% students answered "dream," "water" and "time" as their candidates. Some semantic association is strongly influenced by the user's native language. In the same way, Japanese students tend to look for the expression " His reputation is damaged," and "corrupt" under DOWN, because in Japanese "the reputation goes down" is a very idiomatic expression and "corrupt" is remembered as a Kanji compound daraku (fall+down).

Conclusion

Since it is almost impossible to make the semantic categorization perfect and also it is inconceivable to force the users to abandon their language specific association, or to learn the Activator's look-up system thoroughly, what we need is a comprehensive mutual look-up system and the index in the user's native language.

If some manners of killing are grouped under BREATHE, and not under KILL, there should be a cross reference which indicates the other ones. Then fizzle out, peter out, wear off, fade away, die away can be easily reached and we can find the expressions defendant, prosecution, jury, attorney, plaintiff, etc. just by looking up the word "court." These links to connect similar or related expressions should be much more powerful to make consulting this dictionary more interesting and fruitful.

The index in Japanese might make this dictionary even thicker, but it will solve all the frustration in trying to think of a good Key Word. As shown in Table 2, it was comparatively easy for some words such as "burn" in #3, "feverish" in #4, "submerge" in #6 or "talkative" in #9. About half students got what they wanted. But for some other words it was almost impossible to think of good Key Words, while the students knew exactly what they were looking for in Japanese. Longman Essential Activator, a reduced and more friendly-looking version of this dictionary was recently published. It has a list of headwords at the back instead of in the body of the dictionary. In the same way, Activator's Japanese index could be arranged at the back of the dictionary alphabetically, mixed up with the headword and Key Word list.

Finally, the problem of all the page-turning would be solved by making this dictionary electronic, like one on a CD-ROM. Then the thickness or heaviness of the dictionary would be nothing but an advantage and it could include more example sentences or more usage or grammatical explanations if necessary. We wouldn't have to worry about the physical size of the dictionary.

It might take another few years of research and a lot of manpower to realize all which was suggested here, especially because if Japanese students need a Japanese index, students speaking other languages need an index in other languages too. But I'm sure then all students will fall in love with this dictionary.

References:

Miyai, Shoji, Hiroko Saito, Rumi Takahashi and Kazuo Dohi. An analysis of the Longman Language Activator: The World's First Production Dictionary, LEXICON 25, 26-76, 1995.

Tomioka, Taeko. On the Longman Language Activator and its Productivity, Sanno Junior College Journal 29, 67-81, 1996.